Biological Conservation 159 (2013) 63–72

Contents lists available at SciVerse ScienceDirect

Biological Conservation

journal homepage: www.elsevier .com/ locate /biocon

Can satellite-based night lights be used for conservation? The case ofnesting sea turtles in the Mediterranean

0006-3207/$ - see front matter Crown Copyright � 2012 Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.biocon.2012.11.004

⇑ Corresponding author at: ARC Centre of Excellence for Environmental Decisions,School of Biological Sciences, The University of Queensland, Brisbane, Queensland4072, Australia. Tel.: +61 972 2 6585714.

E-mail addresses: tessa.mazor@uqconnect.edu.au (T. Mazor), noamlevin@msc-c.huji.ac.il (N. Levin), h.possingham@uq.edu.au (H.P. Possingham), yaniv@npa.org.il(Y. Levy), ducciorocchini@gmail.com (D. Rocchini), Anthony.Richardson@csiro.au(A.J. Richardson), salit.kark@gmail.com (S. Kark).

Tessa Mazor a,b,⇑, Noam Levin c, Hugh P. Possingham a, Yaniv Levy d, Duccio Rocchini e,Anthony J. Richardson f, Salit Kark a,b

a ARC Centre of Excellence for Environmental Decisions, School of Biological Sciences, The University of Queensland, Brisbane, Queensland 4072, Australiab The Biodiversity Research Group, Department of Evolution, Ecology and Behaviour, The Silberman Institute of Life Sciences, Hebrew University of Jerusalem, Jerusalem 91904, Israelc Department of Geography, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israeld Israel’s Sea Turtle Rescue Centre, Nature & Parks Authority. Mevoot Yam, P.O.B. 1174, Mikhmoret 40297, Israele Edmund Mach Foundation, Research and Innovation Centre, Department of Biodiversity and Molecular Ecology, GIS and Remote Sensing Unit, Via Mach 1, 38010,San Michele all’Adige (TN), Italyf School of Mathematics and Physics, The University of Queensland, Brisbane, Queensland 4072, Australia

a r t i c l e i n f o

Article history:Received 3 June 2012Received in revised form 22 October 2012Accepted 4 November 2012Available online 21 January 2013

Keywords:Artificial night lightsCaretta carettaChelonia mydasCoastal conservationSatellite imagerySea turtle conservation

a b s t r a c t

Artificial night lights pose a major threat to multiple species. However, this threat is often disregarded inconservation management and action because it is difficult to quantify its effect. Increasing availability ofhigh spatial-resolution satellite images may enable us to better incorporate this threat into future work,particularly in highly modified ecosystems such as the coastal zone. In this study we examine the poten-tial of satellite night light imagery to predict the distribution of the endangered loggerhead (Caretta caret-ta) and green (Chelonia mydas) sea turtle nests in the eastern Mediterranean coastline. Using remotesensing tools and high resolution data derived from the SAC-C satellite and the International Space Sta-tion, we examined the relationship between the long term spatial patterns of sea turtle nests and theintensity of night lights along Israel’s entire Mediterranean coastline. We found that sea turtles nestsare negatively related to night light intensity and are concentrated in darker sections along the coast.Our resulting GLMs showed that night lights were a significant factor for explaining the distribution ofsea turtle nests. Other significant variables included: cliff presence, human population density and infra-structure. This study is one of the first to show that night lights estimated with satellite-based imagerycan be used to help explain sea turtle nesting activity at a detailed resolution over large areas. Thisapproach can facilitate the management of species affected by night lights, and will be particularly usefulin areas that are inaccessible or where broad-scale prioritization of conservation action is required.

Crown Copyright � 2012 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Coastal zones are experiencing rapid population growth aroundthe world (Turner et al., 1996) and attract increasing levels of tour-ism, trade and development (Shi and Singh, 2003; Stancheva,2010). These anthropogenic pressures threaten biodiversity inthe coastal environment, affecting the dynamics of flora and faunapopulations and ecosystem processes (Chapin et al., 2000; Crainet al., 2009). While the effects of some human-caused threats havebeen examined in detail, our understanding of the consequences of

artificial night lights on biodiversity in coastal areas, which haverapidly increased in both spatial extent and intensity in recent dec-ades, remains limited (Longcore and Rich, 2004).

Researchers have studied the effect of night lights on species formany years (Longcore and Rich, 2004). Previous studies exploringthe impact of artificial lights on organisms were mainly conductedby ecologists studying species of birds (e.g. Longcore, 2010), seaturtles (e.g. Lorne and Salmon, 2007), bats (e.g. Jung and Kalko,2010) and freshwater fish (e.g. McConnell et al., 2010). Resultsfrom these studies demonstrate that night lights can attract, repel,and disorientate organisms in their natural settings. These reac-tions can further alter behavioral patterns such as reproduction,foraging, migration, communication and predator–prey relation-ships (Longcore and Rich, 2004). Such studies provide evidencethat artificial lights often have adverse effects on organisms (Sal-mon 2003; Bird et al., 2004; Longcore and Rich, 2004; Bourgeoiset al., 2009; Kempenaers et al., 2010; Longcore, 2010).

64 T. Mazor et al. / Biological Conservation 159 (2013) 63–72

The threats of artificial night lights to biodiversity are rarely ex-plored at a broad spatial scale. Previous studies were predomi-nantly conducted at a local scale in field or laboratory settings(Witherington and Bjorndal, 1991; Salmon et al., 1995b; Grigioneand Mrykalo, 2004). However, broader, regional spatial patternsof activities and processes that threaten the existence of speciesare important to examine, especially when management practicesare applied at larger spatial scales, as is often the case in regionalconservation planning for large marine and terrestrial mammalsand reptiles (Watzold et al., 2006). Today, with our improved abil-ity to estimate anthropogenic pressures and activities from ad-vanced sources such as satellite imagery and remote sensing, weare able explore the impact of human-threats on species at variousscales (Kerr and Ostrovsky, 2003).

Few studies have used satellite night light data for the assess-ment of threats and impacts on species, biological or environmen-tal factors. Of the limited studies, night light imagery has beenused in conservation to derive an index for environmental sustain-ability (Sutton, 2003), has been used to explore the temporal im-pact of light pollution on marine ecosystems (Aubrecht et al.,2010a) and has been incorporated into the management of pro-tected areas (Aubrecht et al., 2010b). However, the effect of artifi-cial light sources and the night environment has largely beenneglected in reserve system or corridor designs (Bird et al., 2004;Longcore and Rich, 2004). No studies, as far as we are aware, haveexplicitly examined the potential of using satellite night lightimagery as a tool for examining the distribution of sea turtle nestsand its further conservation application.

1.1. Sea turtles – threats and factors affecting nesting patterns

Sea turtle species Caretta caretta (Linneaus, 1758, loggerheadturtle) and Chelonia mydas (Linneaus, 1758, green turtle) are glob-ally endangered (Calase and Margaritoulis, 2010). Their worldwideconservation status underlines the importance of understandingfactors that influence their distribution and vulnerability. Sea tur-tles display philopatry, where nesting turtles return to their origi-nal place of birth (Carr, 1975; Bowen et al., 1994). This behavior isknown to operate at a relatively coarse regional scale �10 km–50 km (Miller et al., 2003) and factors that drive nesting sea turtleswithin this coarse spatial-scale are poorly understood (Weisham-pel et al., 2003; Garcon et al., 2009).

One important factor that is known to affect sea turtle behavioris the presence of night lights. Ecologists have found artificial lightsdisrupt sea turtle behavior in two ways. First, night lights reducethe ability of sea turtle hatchlings to find the sea. Hatchlings areeither attracted to the artificial light source or are disorientated(Salmon, 2003; Tuxbury and Salmon, 2005; Lorne and Salmon,2007; Kawamura et al., 2009). Disoriented turtle hatchlings mayfail to find the sea, thereby reducing population viability (Lorneand Salmon, 2007; McConnell et al., 2010).

Second, there is the poorly understood phenomenon of artificialbeach-front lighting preventing turtles from nesting. Nesting fe-males of C. caretta and C. mydas are deterred by artificial lighting(Witherington, 1992; Salmon et al., 1995b; Witherington and Mar-tin, 2000; Bourgeois et al., 2009). The repellent effect could be dosedependent so that highly lit areas deter all nesting and poorly litareas have a minor impact (Margaritoulis, 1985; Witherington,1992). Most of these studies are on beach sites along the coast ofFlorida (Salmon et al., 1995b; Witherington and Martin, 2000; Sal-mon, 2003; Weishampel et al., 2006; Aubrecht et al., 2010a). Seaturtle researchers along the coast of the Mediterranean Sea seldominvestigate this relationship (Kaska et al., 2003; Aureggi et al.,2005) and very few studies have explored this issue at a regionalor broad spatial scale. Overall, the relationship between nightlights and its effect on sea turtle nesting is poorly understood.

Previous studies found that sea turtles nest in non-random pat-terns and their selection of nest site is influenced by specific factors(Mellanby et al., 1998; Weishampel et al., 2003). Besides nightlights, variables that are considered to influence sea turtle nestinginclude: beach dimensions (Kikukawa et al., 1996; Mazaris et al.,2006), beach slope (Wood and Bjorndal, 2000) sand characteristics(Le Vin et al., 1998; Kikukawa et al., 1999), beach nourishment(Brock et al., 2009), climate change (Van Houtan and Halley,2011), predation (Leighton et al., 2011), human settlements(Kikukawa et al., 1996) and coastal development such as seawalls(Rizkalla and Savage, 2011). Understanding the impact of thesevariables on sea turtle nesting is important for setting spatial con-servation priorities (Moilanen et al., 2009).

In this paper we investigate whether night lights, as quantifiedusing space-borne images, can be used to help predict the distribu-tion of sea turtle nests and we discuss the potential application ofthis tool in future conservation applications. The major questionswe test in this study are:

(1) Can night lights derived from satellite imagery help usexplain the distribution of sea turtle nests?

(2) Do night lights remain important at predicting sea turtlenest activity when considering additional anthropogenicand environmental variables?

2. Materials and methods

2.1. Study area

The Mediterranean Sea coastline of Israel is �190 km long andhas a north–south orientation (with the exception of the Carmeland Haifa Bay; Schattner, 1967; Fig. 1). The overall width of bea-ches in Israel is between 20 and 100 m, with wider areas at rivermouths. Israel’s southern beaches (south of Tel Aviv) are character-ized by relatively wider, sandy beaches (compared with northernbeaches) with transverse sand dune fields, which have formed be-hind the shore in the past 1000 years (Schattner, 1967; Tsoar,2000). In comparison, northern beaches are generally narrowerand bordered by aeolionite (kurkar) cliffs. There are 32 rivers andephemeral streams that flow through this coastal stretch into thesea (Lichter et al., 2010) and tidal movements in Israel are limitedto a range of 15–40 cm (Lichter et al., 2010).

Rectangular spatial units along the Israeli coastline were de-signed to examine the relationship between turtle nesting sites,night lights and associated anthropogenic and environmental fac-tors. A buffer of 500 m to the east and west of the coastline wasconstructed and 336 spatial units of 1 � 0.5 km were positionedin this space. The buffer was chosen to allow for longitudinal loca-tion errors, as sea turtle nest surveyors sometimes reported onlythe latitudes. The dimensions of the spatial unit were based onthe resolution of available night light imagery and expert adviceregarding nesting turtle behavior.

2.2. Sea turtle data

Sea turtle data for this study were provided by Israel’s NationalParks Authority (NPA). We used nesting data of the two sea turtlespecies, C. caretta and C. mydas, which nest on the Mediterraneanbeaches of Israel (Kuller, 1999; Levy, 2003). The annual numberof sea turtle nests have been increasing exponential within the pasttwo decades, however specific reasons for their increase are un-known (Levy, 2011; see Appendix Fig. A1). Sea turtle surveys alongthe entire coast of Israel were performed by Israel’s Nature andParks Authority since 1993, during the turtle nesting season fromMay to August. At the start of the nesting season (May), surveyswere conducted two or three times a week. During peak season

Fig. 1. Map showing the study area along the Mediterranean coast of Israel, using the Israel Transverse Mercator Grid. (a) Total number of sea turtle nests summed from 1993to 2011 within each spatial unit (1 � 0.5 km) along the coast of Israel; (b) sea turtle nest occupancy (presence/absence) was summed from 1993 to 2011 within each spatialunit. Israel’s location within the Mediterranean basin is displayed at the bottom. The map was created with ESRI (2011) ArcGIS, Coastline: Survey of Israel, Turtle data: IsraelNature and Parks Authority.

T. Mazor et al. / Biological Conservation 159 (2013) 63–72 65

66 T. Mazor et al. / Biological Conservation 159 (2013) 63–72

(June–July), beaches were surveyed daily. Towards the end of theseason (August), surveys were performed twice a week. For surveypurposes, the Mediterranean coast of Israel was divided equallyinto seven survey sections. Beach sections from Herzliya to Tel Aviv(�8 km) were not surveyed due to high human population densityand development.

The beach sections were scanned at sunrise by Israel’s Natureand Parks Authority rangers along with trained volunteers. Surveyswere conducted with 4WD vehicles driven close to the water edge,with a minimum of two people searching from the windows. Turtlenests were identified by the sand tracks that the female turtleleaves behind after laying her eggs. The two turtle species can eas-ily be identified via their large and unique imprints, nest depth andposition on the sand. The nest position was recorded via GarminGPS units. Turtle tracks that did not result in a nest (false crawl),but seem to clearly be a nesting attempt were also recorded.Hatchling emergence or success was not systematically recordedover the years.

We examined and mapped the turtle nest data using ArcGIS(ESRI, 2011). We combined the two sea turtle species togetherdue to their related choice of nesting beaches (Broderick and God-ley, 1996; Weishampel et al., 2003) and the low number of C. my-das turtle nests in our study (0.8% of all nests). We used twovariables derived from the turtle nest surveys: (1) the total numberof nests found in each spatial unit summed over 19 years (1993–2011; Fig. 1a); (2) the occupancy (presence/absence) status of eachspatial unit for turtle nests in each year and then summed over a19 year period (1993–2011) – this will be referred to as turtle nestpersistence (Fig. 1b). This was performed to limit influences fromindividual years (Fig. A1). When the total number of turtle nestswas summed per spatial unit for this time frame, there was a meanof 9.63 ± 15.5, a median of 3.5 and a range from 0 to 169 individualturtle nests. Twenty-six percent of the surveyed spatial units in ourstudy had no turtle nests (absences).

2.3. Night light data

Two satellite images of the Israel coastline were used for thisstudy, SAC-C (2007; 300 m) and ISS (2003; 60 m). We used a2007 satellite image from Argentine’s Space Agency (CONAE,2007) acquired by the High Sensitivity Technological Camera(HSTC) onboard the SAC-C satellite launched in 2000 (Fig. 2a). Thisimage showed night lights at a spatial resolution of 300 m (Colombet al., 2003) for the entire Israeli coastline. The SAC-C image under-went an inverse Fourier transformation to remove striping effects,using Idrisi Taiga (Clark Labs, 2010; Levin and Duke, 2012). Oursecond image, ISS, was from astronaut photography onboard theInternational Space Station (ISS mission 6). Imagery was obtainedvia Kodad DSC 760 camera at a resolution of 60 m in 2003 (ImageScience and Analysis Laboratory, 2003). The spatial extent of thisimage did not cover the entire Israeli coastline (missing data be-yond Haifa) but was included due to the difficulty of obtaining highspatial resolution satellite images which covers the entire coastlineof Israel. Night light data for 286 of the 336 spatial units were cov-ered by the ISS image (Fig. 2b). For both satellite images we deter-mined an average pixel brightness value for each spatial unit withArcGIS tools (ESRI, 2011).

2.4. Other explanatory variables

In addition to testing the importance of night lights at predict-ing turtle nesting patterns, we examined the effect of 21 additionalvariables that were hypothesized to affect sea turtle nesting andwhich were available for the full study region. These variables weredivided into two groups; anthropogenic and environmental (seeTable 1 for the full list of variables tested).

2.5. Statistical analysis

Our statistical analysis was designed to address our two majorresearch questions;

2.5.1. Satellite night lights and sea turtle nestsWe tested the ability of the two night light images to explain

turtle nest distribution along the coast of Israel. Spearman’s rankcorrelation coefficients were used to test for associations betweenturtle nest distribution and the average pixel values derived fromthe two night light images. To test our hypothesis that turtles pre-fer nesting in darker areas, we split our data into three night lightintensity groups based on pixel values (high, moderate and low –each group with an equal number of spatial units) from both satel-lite images. The three groups were compared via the non-paramet-ric Kruskal–Wallis one-way analysis of variance conducted in Rsoftware (R Development Core Team, 2011). Quantile regressionwas used to further explore the relationship between sea turtlenests and night lights along the entire Israel coastline using theSAC-C image. Quantile regression was performed using the Rquantreg package (Koenker, 2007) with an exponential fit andbootstrapping for residuals.

2.5.2. The importance of satellite night lightsHere we examined the importance of night lights when consid-

ering other variables which may influence sea turtle nest distribu-tion. We also aimed to construct models that predict: (1) the totalnumber of nests per spatial unit and (2) turtle nest persistence, forthe entire Israeli coastline with night lights (using the SAC-C image)and 21 broad scale explanatory variables (Table 1). We used gener-alized linear modeling (GLM) in R. GLMs simultaneously explorewhich variables and/or their interactions explain the highestamount of variability in turtle nest distribution. Prior to beginningthe modeling procedure we tested for collinearity among theexplanatory variables using Spearman rank correlations coefficientand Variance Inflation Factors (VIFs). We used a cut-off value of 3 forremoving collinearity from the resulting VIFs (Zuur et al., 2007), and±0.5 for Spearman’s rank correlations coefficients between pairs ofvariables (Booth et al., 1994). For this analysis we used GLMs witha Poisson distribution, detected overdispersion and corrected thestandard errors using quasi-GLMs (Zuur et al., 2009). Due to devia-tions in the coastline, the area of each spatial unit was not constantand therefore we performed our models with an offset variable forarea (Zuur et al., 2009). Model simplification was conducted bydropping each explanatory variable in turn and removing the termthat led to the smallest non-significant change in deviance accord-ing to F-tests (using the drop1 command in R; Zuur et al., 2009).Model validation was conducted using the deviance residuals plot-ted against the fitted residuals, explanatory variables and spatialcoordinates. We also tested our raw data and models residuals forspatial auto-correlation using spline correlograms with 95% point-wise bootstrap confidence intervals and a maximum lag distanceof 10 km (Bjørnstad and Falck, 2001; Zuur et al., 2009).

3. Results

3.1. Satellite night lights and sea turtle nests

Night lights from the SAC-C image were negatively correlatedwith the total number of sea turtle nests (Spearman’s rho = �0.31,p = 4.07e�09; Fig. 3a) and nest persistence (Spearman’srho = �0.34, p = 8.12e�11; Fig. 3b) across the Israel coastline. Com-parison of the two satellite images when related to sea turtle nestsindicated that the ISS image with the higher resolution gave onlyslightly more significant results compared to the SAC-C image

Fig. 2. The satellite images used in this study for calculating night lights along the coast of Israel. Major cities are displayed. (a) SAC-C satellite from Argentine’s Space Agency(CONAE, 2007), pixel resolution is 300 m. (b) Image from International Space Station astronaut photography, pixel resolution is 60 m (Image Science and Analysis Laboratory,2003). The map was created with ESRI (2011) ArcGIS.

T. Mazor et al. / Biological Conservation 159 (2013) 63–72 67

(Table 2). We found that the total number of sea turtle nests (Krus-kal–Wallis test, SAC-C p = 4.7e�0, ISS p = 1.01e�06; Fig. 4) and nestpersistence (Kruskal–Wallis test, SAC-C p = 3.24e�08, ISSp = 1.28e�07; Fig. 5) within our spatial units were significantly dif-ferent for the three groups of night light intensity. The mean rankof turtle nest numbers was highest in the low pixel group (meanSAC-C = 202.46; ISS = 173.82), which refers to darker sites, com-pared to the mean of the moderate (mean SAC-C = 169.91;ISS = 147.08) and high (mean SAC-C = 133.13; ISS = 111.42) groupsfor both satellite images. Similarly, for both satellite images themean rank of turtle nest persistence was highest in the low pixelgroup (mean SAC-C = 206.50; ISS = 175.28), compared to moderate(mean SAC-C = 167.87; ISS = 148.40) and high (mean SAC-C = 131.13; ISS = 108.65) groups. Quantile regression showed thatthe 0.5 (median) and 0.75 quantiles were statistically significantfor the relationship between night lights and sea turtle nests alongthe entire coastline of Israel (see Appendix Table A1).

3.2. The importance of satellite night lights

Night lights were found to be a significant explanatory variablefor explaining the sea turtle nesting activity in both of our resultingGLMs (Table 3). Our resulting models were able to predict 18%(pseudo r2) of the total number of sea turtle nests and 32% of seaturtle nest persistence within the spatial units along the entirecoast of Israel. Of the 22 (including night lights) explanatory vari-able used in the modeling process, five variables were consideredimportant for explaining the total number of sea turtle nests with-in our spatial units: night lights (F = 7.60, p = 0.01), cliffs (F = 26.22,p = 5.19e�07), the interaction between human population densityand infrastructure (F = 10.22, p = 1.53e�03) and red sandy clayloam (F = 5.63, p = 0.02). Similar variables were considered signifi-cant for explaining sea turtle nest persistence, three two-wayinteractions made up our final model: the interaction betweenbeach area and human population density (F = 4.91, p = 0.03), night

Table 1Table displaying 21 variables used in this study (in GLM). Four anthropogenic based and 17 environmental variables were used that were suspected to be related to turtle nestingpatterns (� = categorical variable).

Variables Data origin

Anthropogenic basedHuman population

densityPopulation density data was obtained as of 2007 for statistical units as defined by Israel Central Bureau of Statistics (CBS, 2007). As a proxy forestimating the population residing near the beach, each spatial unit was given the population density of the closest municipality divisionalongside the coast

Built-up areas (m) Data for built up areas was available from the Israeli Ministry for Environmental Protection (Kaplan et al., 2006), within each spatial unit (CBS,2007). Built-up areas were calculated by the distance from the coastline (middle of spatial unit) to the closest built up area (m)

Infrastructure (m) To determine the land-use type of the beach we used GIS data supplied by the Society for the Protection of Nature in Israel (SPNI) OpenLandscape Institute (OLI). The distance (m) from the center of each spatial unit to beaches clear of national infrastructure (e.g. ports, roads,electrical grids, and military areas) was measured

Reserves The current areas protected within nature reserves and national parks of Israel were provided by Israel’s Nature and Parks Authority. Thepercentage of each rectangular unit that is protected by a reserve which is either officially declared or approved was calculated using ArcGIS(ESRI, 2011). Reserves that are currently awaiting approval or recently proposed were not taken into consideration

Environmental variablesBeach area We digitized the area of beach (sand area) from Google Earth (2011) satellite imagery, performed at the rectangular unit scale (500 m) in

ArcGIS (ESRI, 2011). We calculated the percentage of the spatial unit’s area which was covered by beachCliffs� We included the presence and absence of cliffs bordering the shoreline of beaches as a categorical variable (1 = cliffs, 0 = no cliff). This data was

provided by the Society for the Protection of Nature in Israel (SPNI) Open Landscape Institute (OLI)Geomorphologic

featuresWe used GIS data from a Geological Survey of Israel for the Ministry of Environment (Zilberman et al., 2006). Fifteen geomorphologic classes(Table A3) were considered in our analysis. We calculated the percentage of each geomorphologic feature within every rectangular unit

Fig. 3. Scatter plot using spatial units (1 � 0.5 km) along the coast of Israel to show relationships between sea turtle nesting activity over a 19 years period (1993–2011) andnight light intensity derived from a satellite image (SAC-C; CONAE, 2007). One outlier was removed from the plot for visualization purposes. (a) Total number of sea turtlenests summed per spatial unit (1 � 0.5 km). (b) Sea turtle nesting persistence (presence/absences) summed over time period for each spatial unit.

Table 2Spearman rank correlation coefficient of night lights (pixel values) from two satellite images with sea turtle nest persistence and the total number of sea turtle nests (summedover 19 years period within 336 spatial units) along the coast of Israel.

Satellite night light image Total number of sea turtle nests Sea turtle nest persistence

Spearman’s rank correlation coefficient p Spearman’s rank correlation coefficient p

SAC-C (entire Israel Mediterranean coast) �0.31 4.07e�09 �0.34 8.12e�11ISS (partial coast) �0.37 7.71e�11 �0.39 6.44e�12SAC-C (partial coast as used in ISS image) �0.35 1.11e�09 �0.38 3.20e�11

68 T. Mazor et al. / Biological Conservation 159 (2013) 63–72

lights and cliffs (F = 4.62, p = 0.03) and human population densityand infrastructure (F = 5.57, p = 0.02; Table 3).

The only explanatory variable showing signs of collinearity withnight lights was built up areas along the coast (Spearman’srho = �0.61) however this variable was not significant in our mod-els. We also found that the only interaction with night lights wasthe presence of cliffs in our model that explains sea turtle nest per-sistence. No spatial autocorrelation or collinearity (VIFs all below3; Table A2) among our explanatory variables was found and ourmodels met the validation requirements (Fig. A2; Fig. A3).

4. Discussion

This study demonstrates a novel application of satellite nightlight imagery to help predict nesting activity of endangered sea

turtles. While the impact of artificial night lights on biodiversityis often overlooked, we found that the intensity of coastal nightlights derived from satellite-imagery is a significant determinantof sea turtle nest distribution. Results from our GLMs indicatedthat night light intensity remained an important predictor of seaturtle nest distribution when other anthropogenic and environ-mental factors were considered. For endangered species with largescale spatial movement such as sea turtles, where factors thatinfluence their selection of nesting sites are largely unknown,improving our ability to determine their nesting patterns can en-able us to better direct and target our conservation efforts.

This is one of the first studies to explore the relationship be-tween nesting sea turtles and night lights at a regional spatialscale. Our results indicate that the intensity of artificial night lightsalong the Mediterranean coastline of Israel affects sea turtle nest-

(a) (b)

Fig. 4. Box plots of Kruskal–Wallis one-way analysis of variance of three groups of night light intensity; high (well lit areas), moderate, and low (dark areas) related to thetotal number of sea turtle nests occupancy (summed for years 1993–2011) along the coast of Israel. Pixel values of the three groups are in bracket. One outlier was removedfrom the plot for visualization purposes. (a) SAC-C satellite image (CONAE, 2007), (b) ISS satellite image (Image Science and Analysis Laboratory, 2003).

(a) (b)

Fig. 5. Box plots of Kruskal–Wallis one-way analysis of variance of three groups of night light intensity; high (well lit areas), moderate, and low (dark areas) related to seaturtle nest occupancy (presences/absence) frequency (summed for the years 1993–2011) along the coast of Israel. Pixel values of the three groups are in brackets. One outlierwas removed from the plot for visualization purposes. (a) SAC-C satellite image (CONAE, 2007), (b) ISS satellite image (Image Science and Analysis Laboratory, 2003).

T. Mazor et al. / Biological Conservation 159 (2013) 63–72 69

ing patterns, where well lit beaches have lower occurrences ofnesting turtles. These large scale findings are supported by local-scale studies that show nesting is influenced by night light inten-sity (Margaritoulis, 1985; Witherington, 1992). Thus, our broadscale study provides support for the hypothesis that sea turtlesprefer darker beach sites for nesting. By utilizing information de-rived from satellite night light imagery we can explore broaderspatial patterns between species and the night environment whichwere previously spatially restrictive. Our results suggest that nightlights derived from satellite-based images provide a useful tool forassessing broad-scale spatial patterns of sea turtle nest sites.

In addition to artificial night lights, we identified other new andsignificant variables and their interactions that help predict seaturtle nesting activity at a broad spatial scale. The significant pre-dictors found in both our GLMs, besides night lights, were the pres-ence of cliffs (positive effect), human population density (negativeeffect) and infrastructure (negative effect). Although we were lim-ited with the inclusion of explanatory variables due to data avail-ability at this broad scale, we found new and unexplored

explanatory variables that influence sea turtle nesting. This is thefirst study to find that the presence of coastal cliffs have an impor-tant positive influence on sea turtle nests. Findings by Kikukawaet al. (1999) indicated that beach height is an important variable,and Salmon et al. (1995a) found a positive correlation with tall ob-jects along the shoreline, however to our knowledge, no studieshave explicitly explored the effect of cliffs. While cliffs were a po-sitive effect on sea turtle nests in our study, we suggest that theremay be negative effects in some countries with large tidal rangesor areas where sea levels are beginning to rise (Fish et al., 2005).In such areas the presences of cliffs may cause a barrier for nestingturtles, where the landward movements of nesting turtles are re-stricted, thus a potential cause of nest destruction by sea waterinundation (Fish et al., 2005). We recommend further investigationof other beaches with cliffs around the Mediterranean to betterunderstand the effect that coastal cliffs have on sea turtle nestsand its further application for conservation. Hence, at this broadscale we were able to identify variables that influence sea turtlenesting, which is particularly important to consider in conserva-

Table 3Minimum adequate quasi-Poisson GLM to explain sea turtle nest persistence and the total number of sea turtle nests (between 1993 and 2011) within spatial units along theentire coastline of Israel. See Table 1 for details regarding explanatory variables. Interactions between explanatory variables are marked with a cross. Rows with no values signifyexplanatory variables that were eliminated within the modeling process and did not contribute to the final model.

Explanatory variable Total number of nests Nest persistence

Coefficient SE t p df F p Coefficient SE t p df F p

Night lights (SAC-C image) – negativeexponential

3.34e+10 1.79e+10 1.87 0.06 1 7.60 0.01** 6.39e+10 9.60e+09 6.66 1.18e�10***

Cliffs 8.16e�01 2.30e�01 3.54 4.56e�04*** 1 26.22 5.19e�07*** 1.09e+00 1.67e�01 6.52 2.64e�10***

Infrastructure �2.44e�04 1.31e�04 �1.87 0.06 �3.88e�04 9.03e�05 �4.30 2.30e�05***

Human population density �4.06e�05 3.63e�05 �1.12 0.26 �9.10e�05 3.70e�05 �2.46 0.01*

Beach area 1.70e�02 7.81e�03 2.17 0.03*

Beach area � human population density 1.62e�05 7.57e�06 2.14 0.03* 1 4.91 0.03*

Night lights (neg exp) � cliffs �5.73e+10 2.85e+10 �2.01 0.04* 1 4.62 0.03*

Human population density � infrastructure �5.47e�07 4.96e�07 �1.10 0.27 1 10.22 1.53e�03** �2.80e�07 1.81e�07 �1.54 0.12 1 5.57 0.02*

Red sandy clay loam (Geo_2) �1.8e�02 1.28e�02 �1.46 0.15 1 5.63 0.02*

* Statistical significance – 0.05.** Statistical significance – 0.01.*** Statistical significance – 0.001.

70 T. Mazor et al. / Biological Conservation 159 (2013) 63–72

tion management when very little is known about their spatialdistribution.

Night lights and cliffs as individual components have an impor-tant effect on sea turtle nests and combined have an important po-sitive interaction effect (Table 3). This is exemplified by the case ofNetanya (Fig. 2), a coastal city in Israel where beaches have a highnumber of sea turtle nests, shoreline cliffs and bright night lights.This interaction should be further explored in small-scale fieldstudies to understand the nature of this relationship and the im-pact that cliffs near coastal cities exhibit on nesting sea turtles.Beach areas with bright night lights and beach cliffs may be primeareas to focus conservation efforts for the recovery of nesting seaturtle populations.

Anthropogenic based variables may be useful for predictingspecies distribution and activity within highly modified environ-ments such as the coastal zone. In previous studies at local scales,environmental variables have been predominantly used for deter-mining sea turtle nesting activity (Wood and Bjorndal, 2000; Kara-vas et al., 2005; Mazaris et al., 2006). However, findings from ourstudy suggest that human based variables were important. Otherstudies which have included human based variables have alsofound that sea turtle nests were negatively influenced by such fac-tors. For example, Weishampel et al. (2003) found that nests ofgreen and loggerhead sea turtles increased as the density of humandevelopment was lower along beaches in east Florida. A multipleregression approach by Kikukawa et al. (1999) also found that log-gerhead sea turtle nests in Okinawajima, Japan, significantly in-creased with distance from human settlements. We suggest thattoday with the increasing number of anthropogenic threats onthe coastal environment that inclusion of human based factorsmay serve as helpful predictors of sea turtle nesting patterns orother coastal species.

Artificial night lights may pose a greater threat to sea turtlenests compared with other anthropogenic threats. Our GLM resultsshowed that night lights were more significant at explaining seaturtle nests distribution then other anthropogenic threats such asthe human population density, infrastructure and built up areas.Unlike these other variables, night lights account for the presenceof most human night time activity, including beach side restau-rants, shopping districts, ports and residential areas. Interestingly,we also found that higher resolution satellite night light imagery,comparison between the ISS and SCC-C images, was better relatedto sea turtle nesting patterns (Table 2). Thus, the threat of nightlights on sea turtle nesting, while evident from laboratory andsmall-scale field experiments (Witherington, 1992; Salmon et al.,1995b) can also be explored with the use of high resolution satel-lite imagery.

To date, very few explanatory variables and models have beenidentified which can aid our understanding of nesting patterns ofendangered sea turtle species (Garcon et al., 2009). Clearly thereare additional unknown factors which affect sea turtle nest distri-bution. Our resulting models were able to explain 18% and 32% ofturtle nest variance. These values suggest that there are other fac-tors which contribute to predicting sea turtle nest distribution.Other contributing factors could be related to the hypothesis thatsea turtles use multiple environmental factors/cues with thresh-olds to reach before choosing a nesting site (Wood and Bjorndal,2000; Mazaris et al., 2006). Alternatively, these factors could bedue to recently explored climatic factors, predation, other anthro-pogenic threats, interactions among variables (Leighton et al.,2011; Rizkalla and Savage, 2011; Van Houtan and Halley, 2011)or small scale environmental conditions that are not found at thislarge scale (Wood and Bjorndal, 2000). Thus, with the little knowl-edge we have on sea turtle nesting patterns, combined with theirendangered status, we propose that satellite night light imagerymay be a useful tool for the prediction of sea turtle nest distribu-tion at a broad spatial scale and recommend its incorporation intofuture studies.

4.1. Conservation implications

The advancements in spatial analysis and applications (Senet al., 2006) continually allow us to consider new techniques andmethods to explore and predict species assemblages and patternsat broader spatial scales with higher resolution (Kerr and Ostrov-sky, 2003; Turner et al., 2003). In recent years studies have beenquantifying biodiversity with remote sensing tools and satelliteimagery (Levin et al., 2007; Lahoz-Monfort et al., 2010; Rocchiniet al., 2010; Bradter et al., 2011). While such tools and methodscannot replace field work at smaller scales, they can serve as usefultools for exploring larger spatial-scales. In particular circumstanceswhere field work locations are inaccessible or spatial extents aretoo large, remote sensing can provide us with the best knowledgeat hand. Further research therefore, should be conducted withthese tools at broader spatial scales and regional levels in orderto advance our understanding of species habitat selection, move-ment and threats.

Predicting species habitats, movements and identifying theirthreats can greatly aid conservation decisions, which are oftenmade with relatively sparse information (Pressey, 2004). Whilethis study examines nesting sea turtles, the same methodologycan be applied to other species that are disturbed by artificial nightlights. For such species, we propose that satellite night light imag-ery can be incorporated into conservation planning in order to mit-

T. Mazor et al. / Biological Conservation 159 (2013) 63–72 71

igate the threat of night lights when selecting priority conservationareas or reserves. This approach is especially relevant for rare andendangered species such as sea turtles, for which there is a limitedtime to act in the face of increasing human-pressures and whereaction is needed at broad scales.

Acknowledgments

We would like to thank the Israel Nature and Parks Authority’srangers and the many people who collected data and sent reportson turtle nest site locations over the past two decades. We thankthe Society for the Protection of Nature in Israel (SPNI) Open Land-scape Institute (OLI) and Israel Ministry of Environmental Protec-tion for providing GIS data. T.M. gratefully acknowledges thefinancial support of the Australia–Israel Scientific ExchangeFoundation.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.biocon.2012.11.004.

References

Aubrecht, C., Elvidge, C.D., Ziskin, D., Rodrigues, P., Gil, A., 2010a. Observing stress ofartificial night lighting on marine ecosystems – a remote sensing applicationstudy. In: Wagner, W., Székely, B. (Eds.), ISPRS TC VII Symposium – 100 YearsISPRS. IAPRS, Vienna, pp. 41–46.

Aubrecht, C., Jaiteh, M., de Sherbinin, A., 2010b. Global Assessment of Light PollutionImpact on Protected Areas. CIESIN/AIT Working Paper. Palisades, NY, USA:CIESIN and NASA SEDAC, The Earth Institute at Columbia University.

Aureggi, M., Rizk, C., Venizelos, L., 2005. Survey on sea turtle nesting activity SouthLebanon. MEDASSET and MEDWESTCOAST. <www.medasset.org> (accessed01.2012).

Bird, B., Branch, L., Miller, D., 2004. Effects of coastal lighting on foraging behaviourof beach mice. Conservation Biology 18, 1435–1439.

Bjørnstad, O.N., Falck, W., 2001. Nonparametric spatial covariance functions:estimation and testing. Environmental and Ecological Statistics 8, 53–70.

Booth, G.D., Niccolucci, M.J., Schuster, E.G., 1994. Identifying proxy sets in multiplelinear regression: an aid to better coefficient interpretation. Research paperINT-470. United States Department of Agriculture, Forest Service, Ogden.

Bourgeois, S., Gilot-Fromont, E., Villefont, A., Boussamba, F., Deem, S., 2009.Influence of artificial lights, logs and erosion on leatherback sea turtlehatchling orientation at Pongara National Park. Gabon Biological Conservation142, 85–93.

Bowen, B.W., Kamezaki, N., Limpus, C.J., Hughes, G.R., Meylan, A.B., Avise, J.C., 1994.Global phylogeography of the loggerhead turtle (Caretta caretta) as indicated bymitochrondrial DNA haplotypes. Evolution 48, 1820–1828.

Bradter, U., Thom, T., Altringham, J.D., Kunin, W.E., Benton, T.G., 2011. Prediction ofNational Vegetation Classification communities in the British uplands usingenvironmental data at multiple spatial scales, aerial images and the classifierrandom forest. Journal of Applied Ecology 48, 1057–1065.

Brock, K.A., Reece, J.S., Ehrhart, L.M., 2009. The effects of artificial beachnourishment on marine turtles: differences between loggerhead and greenturtles. Restoration Ecology 17, 297–307.

Broderick, A.C., Godley, B.J., 1996. Population and nesting ecology of the greenturtle, Chelonia mydas, and the loggerhead turtle, Caretta caretta, in northernCyprus. Zoology in the Middle East 13, 27–46.

Calase, P., Margaritoulis, D. (Eds.), 2010. Sea Turtles in the Mediterranean:Distribution, Threats and Conservation Priorities. IUCN, Gland, Switzerland.

Carr, A.F., 1975. The Ascension Island green turtle colony. Copeia 1975, 547–555.CBS, 2007. Demographics of Israel, Central Bureau of Statistics Israel 2011. <http://

www1.cbs.gov.il/reader/> (accessed 10.2011).Chapin III, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds, H.L.,

Hooper, D.U., Lavorel, S., Sala, O.E., Hobbie, S.E., Mack, M.C., Díaz, S., 2000.Consequences of changing biodiversity. Nature 405, 234–242.

Clark Labs, 2010. IDRISI Taiga 16.05. 950 Main Street, Worcester MA 01610-1477,USA: Clark University.

Colomb, F.R., Alonso, C., Hofmann, C., Nollmann, I., 2003. SAC-C mission, an exampleof international cooperation. Advances in Space Research 34, 2194–2199.

CONAE, 2007. National Space Activities Commission, Satellite – SAC-C, BuenosAires, Argentine.

Crain, C.M., Halpern, B.S., Beck, M.W., Kappel, C.V., 2009. Understanding andmanaging human threats to the coastal marine environment. Annals of the NewYork Academy of Sciences 1162, 39–62.

ESRI, 2011. ArcGIS Desktop: Release 10. Redlands, CA: Environmental SystemsResearch Institute.

Fish, M.R., Cote, I.M., Gill, J.A., Jones, A.P., Renshoff, S., Watkinson, A.R., 2005.Predicting the impact of sea-level rise on caribbean sea turtle nesting habitat.Conservation Biology 19, 482–491.

Garcon, J.S., Grech, A., Moloney, J., Hamann, M., 2009. Relative exposure index: animportant factor in sea turtle nesting distribution. Aquatic Conservation:Marine and Freshwater Ecosystems 20, 140–149.

Google Earth, 2011. Israel coast line, Data SIO, NOAA US Navy NGA, GEBCO. <http://www.google.com/earth/index.html> (accessed 11.2011).

Grigione, M.M., Mrykalo, R., 2004. Effects of artificial night lighting onendangered ocelots and nocturnal prey along the United States–Mexicoborder: a literature review and hypotheses of potential impacts. UrbanEcosystems 7, 65–77.

Image Science and Analysis Laboratory, 2003. NASA-Johnson Space Centre. TheGateway to Astronaut Photography of Earth. <http://eol.jsc.nasa.gov/scripts/sseop/photo.pl?mission=ISS007&roll=E&frame=16433>02/05/201210:24:31>(accessed 05.2012).

Jung, K., Kalko, E., 2010. Where forest meets urbanization: foraging plasticity ofaerial insectivorous bats in an anthropogenically altered environment. Journalof Mammalogy 91, 144–153.

Kaplan, M., Din, H., Bookwald, S., Dabcheri-Darom, L., 2006. Land Use Patterns in theBuilt-up Areas in 2003 and a Comparative Research 1998–2003. The JerusalemInstitute for Israel Studies and Israel’s Ministry of the Environment (in Hebrew).

Karavas, N., Georghiou, K., Arianoutsou, M., Dimopoulos, D., 2005. Vegetation andsand characteristics influencing nesting activity of Caretta caretta on Sekaniabeach. Biological Conservation 121, 177–188.

Kaska, Y., Baskale, E., Urhan, R., Katilmis, Y., Gidis, M., Sari, F., Sozbilen, D., Canbolat,F., Yilmaz, F., Barlas, M., Ozdemir, N., Ozkul, M., 2003. Natural andanthropogenic factors affecting the nest-site selection of Loggerhead Turtles,Caretta caretta, on Dalaman-Sarigerme beach in South–West Turkey. Zoology inthe Middle East 50, 47–58.

Kawamura, G., Naohara, T., Tanaka, Y., Nishi, T., Anraku, K., 2009. Near-ultravioletradiation guides the emerged hatchlings of loggerhead turtles Caretta caretta(Linnaeus) from a nesting beach to the sea at night. Marine and FreshwaterBehaviour and Physiology 42, 19–30.

Kempenaers, B., Borgstrom, P., Loes, P., Schlicht, E., Valcu, M., 2010. Artificial nightlighting affects dawn song, extra-pair siring success, and lay date in songbirds.Current Biology 20, 1735–1739.

Kikukawa, A., Kamezaki, N., Hirate, K., Ota, H., 1996. Distribution of nesting sites ofsea turtles in Okinawajima and adjacent islands of the central Ryukyus. JapanChelonian Conservation and Biology 2, 99–101.

Kikukawa, A., Kamezaki, N., Ota, K., 1999. Factors affecting nesting beach selectionby loggerhead turtles (Caretta caretta): a multiple regression approach. Journalof Zoology 249, 447–454.

Koenker, R., 2007. quantreg: Quantile Regression, R package version 4.06. <http://www.r-project.org> (accessed 02.2012).

Kuller, Z., 1999. Current status and conservation of marine turtles on theMediterranean Coast of Israel. Marine Turtle Newsletter 86, 3–5.

Kerr, J., Ostrovsky, M., 2003. From space to species: ecological applications forremote sensing. Trends in Ecology and Evolution 18, 299–305.

Lahoz-Monfort, J., Guillera-Arroita, G., Milner-Gulland, E.J., Young, R.P., Nicholson,E., 2010. Satellite imagery as a single source of predictor variables for habitatsuitability modelling: how Landsat can inform the conservation of a criticallyendangered lemur. Journal of Applied Ecology 47, 1094–1102.

Leighton, P., Horrocks, J., Kramer, D., 2011. Predicting nest survival in sea turtles:when and where are eggs most vulnerable to predation? Animal Conservation14, 186–195.

Levin, N., Duke, Y., 2012. High spatial resolution night-time light images fordemographic and socio-economic studies. Remote Sensing of Environment 119,1–10.

Levin, N., Shmida, A., Levanoni, O., Tamari, H., Kark, S., 2007. Predicting mountainplant richness and rarity from space using satellite-derived vegetation indices.Diversity and Distributions 13, 692–703.

Le Vin, D.A., Broderick, A.C., Godley, B.J., 1998. Effects of offshore features on theemergence point of marine turtles in Northern Cyprus. In: Byles, R., Fernandez,Y. (Eds.), Proceedings of the 16th Annual Symposium on Sea Turtle Biologyand Conservation. NOAA Technical, Memorandum NMFS-SEFSC-412,pp. 91–92.

Levy, Y., 2003. Status of Marine Turtles and Conservation efforts along the IsraeliCoastline. In: Seminoff, J.A. (Ed.), Proceedings of the 22nd Annual Symposiumon Sea Turtle Biology and Conservation. NOAA Technical, Memorandum NMFS-SEFSC-503, p. 149.

Levy, Y., 2011. Summary of recovery activity of sea turtles in Israel 2011. Annualreport (in Hebrew). Israel Nature and Parks Authority, Mikhmoret.

Lichter, M., Zviely, D., Klein, M., 2010. Morphological patterns of south easternMediterranean river mouths: the topographic setting of the beach as a forcingfactor. Geomorphology 123, 1–12.

Longcore, T., 2010. Sensory ecology: night lights alter reproductive behaviour ofblue tits. Current Biology 20, 893–895.

Longcore, T., Rich, C., 2004. Ecological light pollution. Frontiers in Ecology and theEnvironment 2, 191–198.

Lorne, K., Salmon, K., 2007. Effects of exposure to artificial lighting on orientation ofhatchling sea turtles on the beach and in the ocean. Endangered SpeciesResearch 3, 23–30.

Margaritoulis, D., 1985. Preliminary observations on the breeding behaviour andecology of Caretta caretta in Zakynthos. Greece Biologia Gallo-Hellenica 10,323–332.

72 T. Mazor et al. / Biological Conservation 159 (2013) 63–72

Mazaris, A.D., Matsinos, Y.G., Margaritoulis, D., 2006. Nest site selection ofloggerhead sea turtles: the case of the island of Zakynthos, W Greece. Journalof Experimental Marine Biology and Ecology 336, 157–162.

McConnell, A., Routledge, R., Connors, B.M., 2010. Effect of artificial light on marineinvertebrate and fish abundance in an area of salmon farming. Marine Ecology-Progress Series 419, 147–156.

Mellanby, R.J., Broderick, A.C., Godley, B.J., 1998. Nest site selection inMediterranean marine turtles at Chelones Bay, Northern Cyprus. In:Proceedings of the 16th Annual Symposium on Sea Turtle Biology andConservation (compilers R. Byles and Y. Fernandez). NOAA Technical,Memorandum NMFS-SEFSC-412, pp. 103–104.

Miller, J.D., Limpus, C.J., Godfrey, M.H., 2003. Nest site selection, oviposition, eggs,development, hatching, and emergence of loggerhead turtles. In: Bolton, A.B.,Witherington, B.E. (Eds.), Loggerhead Sea Turtles. Smithsonian Institution,Washington, DC, pp. 125–143.

Moilanen, A., Possingham, H.P., Polasky, S., 2009. A mathematical classification ofconservation prioritization problems. In: Moilanen, A., Wilson, K.A.,Possingham, H.P. (Eds.), Spatial Conservation Prioritisation: QuantitativeMethods and Computational Tools. Oxford University Press, Oxford, pp. 28–42.

Pressey, R.L., 2004. Conservation planning and biodiversity: assembling the bestdata for the job. Conservation Biology 18, 1677–1681.

R Development Core Team, 2011. R: Version 2.13.0. A Language and Environmentfor Statistical Computing. R Foundation for Statistical Computing, Bristol, UK.<http://www.R-project.org> (accessed 01.2012).

Rizkalla, C.E., Savage, A., 2011. Impact of seawalls on loggerhead sea turtle (Carettacaretta) nesting and hatching success. Journal of Coastal Research 27, 166–173.

Rocchini, D., Balkenhol, N., Carter, G., Foody, G., Gillespie, T., He, K., Kark, S., Levin, N.,Lucas, K., Luoto, M., Nagendra, H., Oldeland, J., Ricotta, C., Southworth, J.,Neteler, M., 2010. Remotely sensed spectral heterogeneity as a proxy of speciesdiversity: recent advances and open challenges. Ecological Informatics 5, 318–329.

Salmon, M., Reiners, R., Lavin, C., Wyneken, J., 1995a. Behaviour of loggerhead seaturtles on an urban beach. I. Correlates of nest placement. Journal ofHerpetology 29, 560–567.

Salmon, M., Tolbert, M., Painter, D.P., Goff, M., Reiners, R., 1995b. Behavior ofloggerhead sea turtles on an urban beach. II. Hatchling orientation. Journal ofHerpetology 29, 568–576.

Salmon, M., 2003. Artifical night lighting and sea turtles. Biologist 50, 163–168.Schattner, I., 1967. Geomorphology of the Northern Coast of Israel. Landscape and

Processes: Essays in Geomorphology 49, 2–4.Sen, A., Kim, Y., Caruso, D., Lagerloef, G., Colomb, R., Yueh, S., et al., 2006. Aquarius/-

SAC-D mission overview. Proceedings of SPIE 6361, 63610I–63611I.Shi, H., Singh, A., 2003. Status and interconnections of selected environmental

issues in the global coastal zones. A Journal of the Human Environment 32,145–152.

Stancheva, M., 2010. Human-induced impacts along the coastal zone of Bulgaria. Apressure boom versus environment. Bulgarian Academy of Sciences 63, 137–146.

Sutton, P.C., 2003. An empirical environmental Sustainability index derived solelyfrom nighttime satellite imagery and ecosystem service valuation. Populationand Environment 24, 293–311.

Tsoar, H., 2000. Geomorphology and paleogeography of sand dunes that haveformed the kurkar ridges in the coastal plain of Israel. Israel Journal of EarthSciences 49, 189–196.

Turner, R., Subak, S., Adger, W., 1996. Pressures, trends, and impacts in coastalinteractions between socioeconomic and natural systems. EnvironmentalManagement 20, 159–173.

Turner, W., Spector, S., Gardiner, N., Fladeland, M., Sterling, E., Steininger, M., 2003.Remote sensing for biodiversity science and conservation. Trends in Ecologyand Evolutions 18, 306–314.

Tuxbury, S.M., Salmon, M., 2005. Competitive interactions between artificiallighting and natural cues during sea finding by hatchling marine turtles.Biological Conservation 121, 311–316.

Van Houtan, K., Halley, J., 2011. Long-term climate forcing in loggerhead sea turtlenesting. PloS One 6, e19043.

Watzold, F., Drechsler, M., Armstrong, C., Baumgartner, S., Grimm, V., Huth, A.,Perrings, C., Possingham, H.P., Shogren, J., Skonhoft, A., Verboom-vasiljev, J.,Wissel, C., 2006. Ecological-economic modeling for biodiversity management:potential, pitfalls, and prospects. Conservation Biology 20, 1034–1041.

Weishampel, J.F., Bagley, D.A., Ehrhart, L.M., Rodenbeck, B.L., 2003. Spatiotemporalpatterns of annual sea turtle nesting behaviours along an East Central Floridabeach. Biological Conservation 110, 295–303.

Weishampel, J., Bagley, D.A., Ehrhart, L.M., 2006. Intra-annual loggerhead and greenturtle spatial nesting patterns. Southeastern Naturalist 5, 453–462.

Witherington, B., 1992. Behavioral-responses of nesting sea-turtles to artificiallighting. Herpetologica 48, 31–39.

Witherington, B.E., Bjorndal, K.A., 1991. Influences of artificial lighting on theseaward orientation of hatchling loggerhead turtles Caretta caretta. BiologicalConservation 55, 139–149.

Witherington, B.E., Martin, E.R., 2000. Understanding, Assessing and resolving light-pollution problems on sea turtle nesting beaches (Technical report, TR-2).Florida Marine Research Institute, St. Petersburg, Florida.

Wood, D.W., Bjorndal, K.A., 2000. Relation of temperature, moisture, salinity, andslope to nest site selection in Loggerhead Sea Turtles. Copeia 1, 119–128.

Zilberman, E., Ilani, S., Netzer-Cohen, H., Kalbo, R., 2006. Geomorphologic –lithologic mapping along the coastal plain of Israel (in Hebrew). GeologicalInstitute of the Minister of Energy and Infrastructure Israel, Jerusalem.

Zuur, A.F., Ieno, E.N., Smith, G.M., 2007. Analysing Ecological Data. Springer, NewYork.

Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed effectsmodels and extensions in ecology with R. Springer, New York.